Attention

Element benchmarks

In short

Areas of type Face, Logo or Text are compared with the typical attention share of that element type at a similar size. The reference is not a sector or platform benchmark. Raw shares are comparable within a creative; gaps to the reference are type-specific and should not be ranked across face, logo and text.

What it is

When you give an area of interest the type Face, Logo or Text, Foresight compares it with a reference value: the typical attention share, and typical view time, of that element type at a similar size relative to the image. In the app a chip shows the difference, for example "+3 pts vs Face", and the reference value appears when you hover it.

Custom areas have no element reference. The gaze plot and the attention build-up have no benchmark at all.

What feeds it

The element type is the type you choose for the area. Its size comes from how much of the image your rectangle covers.

How it's built

• Size-aware. Bigger areas naturally capture more attention, so each area is compared with elements of its own type and a similar size relative to the image.
• General reference images. Each element type is compared with general reference images of that type and size: not advertising-specific. They are analysed with the same attention model that analyses your creative.
• Not a sector or platform benchmark. The references are split by element type and size only. They are the same for every sector, creative type and platform. They describe how faces, logos and text typically perform in images in general, not in your category. The app calls it a general benchmark.

The reference is shown for every image area of type Face, Logo or Text, whether or not your analysis has a comparison group, and also when the image is shown inside a platform frame.

Scale and direction

The reference is shown in the same units as your area: attention share in % and view time in seconds. The difference chip is in percentage points: "+3 pts" means your share is 3 points above the typical value, not 3% higher.

Above the reference means the element draws more predicted attention than a typical element of its type and size. This is a separate system from the KPI benchmarks: it is not a 0–100 positioning and has no colour tiers.

Image vs video

Images only. There are no areas of interest, and so no element references, for videos.

How to read it

Comparing elements within one creative: use the raw shares. They are all a percentage of the same creative's attention, so you can say "the face draws more attention than the logo here".

Comparing an element with its reference: read the gap per element. "This logo draws more attention than a typical logo of its size" is a valid conclusion.

Comparing across face, logo and text: do not rank elements by their gaps to the reference. Each type is measured against its own yardstick, so "+5 points" for a face and "+5 points" for a logo do not mean the same thing.

Common pitfalls

• The reference is not your industry. It is the same for every sector and platform.
• The type decides the reference. A product given the type Logo is compared with logos; choose types carefully.
• The chip shows points, not a percentage change.
• A gap is not a performance ranking across element types. Do not conclude that text "performs better" than the logo because its gap is larger.